[rebelos]

Of course it matters: (1) at the macro level, fraud is a tax on the whole ecosystem that ultimately yields higher prices for consumers and (2) at the micro level, relatively better ad efficiency (less fraud) can drive business competitiveness (or even viability, in extreme cases).

[chasebank]

In reference to (1) - I'm not so sure this is the case. Competition amongst advertisers will drive up CPM / CPC until the point of margin. The real loser here is the publisher, wasting money on server bandwidth.

As an example, if the max CPA for furniture suppliers is $500, if we eliminate ad fraud, furniture suppliers won't get a cheaper CPA, they'll competitively bid up the CPM / CPC until the point of a $500 CPA once again.

[from]

> The real loser here is the publisher, wasting money on server bandwidth.

The marginal cost of delivering a web page is near 0. In fact, many major publishers frequently buy "traffic" (read: bot clicks) that is designed to bypass the major bot detection firms like Doubleverify, IAS, etc. It makes them a lot of money and bots don't care about the quality of the articles. There's another related category of fraud called "cashout sites" or "ghost sites" where they make a website (usually celebrity news), get approved to put ads on it, then get a bunch of botted ad clicks on it.

> As an example, if the max CPA for furniture suppliers is $500, if we eliminate ad fraud, furniture suppliers won't get a cheaper CPA, they'll competitively bid up the CPM / CPC until the point of a $500 CPA once again.

In every fraud the fraudster siphons dollars at someone else's expense. Those millions of dollars the government seized in the Methbot case necessarily have to have come from somewhere. An ad fraudster makes advertisers pay for views/clicks that never actually reached real people. Paying $1/click in a system with 0 fraud has a higher ROI for advertisers than paying $1/click in a system where 10% of clicks are fake.

[lifeisstillgood]

I think it's who does the tax go to. Fine if FurnitureWorld.com pays its 500 bucks to google and NYT, but at the moment a chuck of that goes to organised crime - frankly the same people taking a slice of credit card fraud that banks see as a cost of business.

Crime and corruption play a huge part in our societies and relatively simple measures can starve them of the cash flow - from decriminalisation to de-anonymising the net. Or something

[reillyse]

(Ignoring the premise that this is an efficient market - which it very much isn't) So ad fraud is a victimless crime? Where is the money that the fraudsters are earning coming from? Who are they taking that money from?

[cdbyr]

It's an interesting question.

If the fraud were evenly spread out, and if all advertisers had the same goal, the equilibrium bid just adjusts for less valuable clicks/interaction/etc (in line with what the earlier commenter mentioned), and the advertisers pay the same amount as in a world without fraud. Fraud isn't evenly spread out though, and advertisers are sometimes unaware, so it probably does hurt them.

The other loss is from honest publishers (think newspaper websites, etc.) - they're having to split payments with fraudsters, even though they're providing all the value to the advertiser. Downstream effects mean the publisher is probably producing less, showing more ads, or using other ways to replace ad income.

[thehappypm]

It really doesn’t matter in practice. It’s all priced in. The actual cost of serving an ad is absolutely tiny, fractions of a penny. Fraud eats into margins, but really not by much. Advertisers are paying for performance and it’s fully expected that a tiny fraction of ads lead to the desired outcome, but that’s why ads are cheap on a per impression basis.